Abstract
The causal role of cigarette smoking in the risk of colorectal neoplasm has been suggested but not established. In a case–control study including 2060 colorectal polyp patients and 3336 polyp-free controls, we evaluated 21 functional genetic variants to construct a tobacco-carcinogen-metabolizing genetic risk score. Data regarding cigarette smoking were obtained through telephone interviews. Cigarette smoking was associated with an elevated risk of both adenomas and hyperplastic polyps. The association with smoking was stronger in participants with a high carcinogen-metabolizing risk score than those with a low risk score. Smoking 30 or more cigarettes per day was associated with a 1.7-fold elevated risk of any polyps (95% confidence interval = 1.3–2.2) among those with a low genetic risk score and 2.9-fold elevated risk (95% confidence interval = 1.8–4.8) among those with a high genetic risk score (P interaction = 0.025). A similar pattern of interaction was observed in analyses conducted separately for those with adenomas only (P interaction = 0.039) and hyperplastic polyps only (P interaction = 0.024). Interaction between carcinogen-metabolizing genetic risk and cigarette smoking was found in relation to high-risk adenomas (P interaction = 0.010) but not low-risk adenomas (P interaction = 0.791). No apparent interaction was found for duration of smoking. This study shows that the association between cigarette smoking and colorectal polyp risk is modified by tobacco-carcinogen-metabolizing polymorphisms, providing support for a causal role of cigarette smoking in the etiology of colorectal tumors.
Introduction
Cigarette smoking has been linked to the risk of colorectal cancer and adenomas (1,2). Colorectal adenomas are precursors of colorectal cancer (3,4); they can develop into colorectal cancer mostly through an adenoma–carcinoma sequence (5). Recent evidence suggests that some hyperplastic polyps (HPP) may develop into cancer via serrated or microsatellite instable pathways (6,7). It is common for an individual to have both synchronous HPP and adenomas (8,9). It has been shown that patients with both synchronous HPP and adenomas are more likely to have recurrence of adenomas than those with either adenomas or HPP (9). Studies have shown cigarette smoking to be more strongly associated with the risk of both synchronous HPP and adenomas and the risk of HPP only than adenomas only. However, the association between genetic variation in genes encoding tobacco-carcinogen-metabolizing enzyme and the risk of colorectal adenoma and HPPs has been inconsistent.
Cigarette smoke contains a variety of carcinogenic compounds, including polycyclic aromatic hydrocarbons, heterocyclic amines, aromatic amines and nitrosamines (10). Most tobacco carcinogens require metabolic activation by Phase I enzymes before they are able to react with DNA. This metabolic activation often is initiated by cytochrome P450 enzymes (CYPs, i.e. CYP1A2 and CYP1B1). These active metabolites are then detoxified by Phase II enzymes, including UDP-glucuronosyltransferases (UGT) and glutathione-S-transferases (GST), and then excreted from the body (10,11). Aryl hydrocarbon receptor (AhR) is an important xenobiotic signaling mediator to enhance the expression of both Phase I and Phase II enzymes (12), which affects tobacco-carcinogen metabolism. Internal doses of tobacco-carcinogen exposure, and the resulting biological effects, are influenced by the balance of enzymes that activate and detoxify tobacco carcinogens. The amount and function of key carcinogen-metabolizing enzymes are determined, in part, by genetic polymorphisms of genes encoding these enzymes. Categorization of tobacco-carcinogen-metabolizing enzyme genotypes provides a tool to classify participants into groups with different internally active carcinogen-exposure levels despite the same external exposure level. In other words, if cigarette smoking is causally associated with colorectal polyp risk, we would expect that this association may be modified by genetic polymorphisms of enzymes involved in the metabolism of tobacco carcinogen. Individuals with a high internal exposure to biologically active tobacco carcinogen, as predicted by both external exposure assessment and tobacco-carcinogen-metabolizing pattern, may be at an elevated risk of polyps.
We used data and DNA samples collected in the Tennessee Colorectal Polyp Study (TCPS) to test this hypothesis. Because tobacco-carcinogen-metabolizing genotypes are established through random assortment during gamete formation, they should be independent of external exposure to tobacco carcinogens and are unlikely to be related to confounding factors. Our study is consistent with Mendelian randomization analysis in studying gene–environment interactions in the risk of diseases, which helps reduce potential biases associated with observational studies.
Participants and methods
Recruitment of study participants
The TCPS is a colonoscopy-based case–control study conducted in Nashville, TN. Detailed methods used in this study have been described elsewhere (13,14). Eligible participants of age 40–75 years were identified from patients scheduled for colonoscopy at an academic medical center (Vanderbilt University Medical Center) and a Veterans Affairs medical center (Tennessee Valley Healthcare System, Nashville, TN) between 1 February 2003 and 26 March 2010. The study was approved by the Institutional Review Boards of all participating institutions.
Excluded from our study were participants with genetic colorectal cancer syndromes or with prior history of inflammatory bowel disease, colorectal adenomas, or any cancer other than non-melanoma skin cancer. Among 10 074 eligible participants, 7330 (72.8%) provided written informed consent, of whom 6331 (86.4% of responders) completed a telephone interview. Based on the colonoscopy and pathologic findings, polyp cases were categorized as follows: (i) cases with HPP only, (ii) cases with adenoma only and (iii) cases with both synchronous adenoma and HPP. Eligible controls were participants who had received a complete colonoscopy reaching the cecum and were found to be polyp free. Twenty-four cases were excluded due to missing data. Among those with completed interviews, 5396 participants were genotyped. The current analyses include 1160 cases with adenoma only, 532 cases with HPP only, 368 cases with both synchronous adenoma and HPP and 3336 polyp-free controls. Based on endoscopic report, advanced adenomas were defined as adenomas with a diameter ≥ 1cm, high-grade dysplasia, or tubulovillous or villous morphology.
Assessment of cigarette smoking and other lifestyle factors
After colonoscopy, trained interviewers conducted a standardized telephone interview to obtain information regarding medication use, demographics, medical history and selected lifestyle factors, including detailed information about cigarette smoking (13). Briefly, regular cigarette smoking (or ever smoking) was defined as smoking at least one cigarette per day for at least 3 months continuously. Former smokers were regular smokers who had stopped for at least 1 year before colonoscopy. Information was collected about age at initiation of cigarette smoking, whether the participant was still smoking regularly, age at which regular smoking stopped, average number of cigarettes smoked per day (either currently or before quitting) and maximum number of cigarettes smoked per day regularly during the participant’s lifetime. Regular alcohol drinking was defined as consuming five or more drinks per week for 12 months continuously. Regular non-steroidal anti-inflammatory drug (NSAID) users were defined as those using NSAIDs at least three times each week for at least 12 months continuously. All cut-points for exposure variables were based on distributions in control participants.
Genetic variant selection and genotyping assays
In this study, we selected 12 key enzymes involved in tobacco-carcinogen metabolism (Table I). We attempted to identify all well-established functional genetic variants involved in the entire tobacco-carcinogen-metabolism pathway based on literature showing functional changes of enzymes by in vitro studies of these variants. Tobacco-carcinogen-metabolizing enzymes were categorized into Phase I and Phase II enzymes (Table I). We identified 27 functional variants of these enzymes, of which 25 were successfully genotyped. Those which failed genotyping were C1095A (3′UTR, rs15561) and T1088A (3′UTR, rs1057126) in NAT1 gene. Four NAT1 single nucleotide polymorphisms (SNPs), that is C97T (R33Stop), C190T (R64W), C620T (T207I) and A752T (D251V), were monomorphic in our population and thus were excluded. As a result, two NAT1 SNPs remained in the analysis, that is NAT1*14 (G560A, rs4986782) and NAT1*15 (C559T, rs5030839). The other 19 genetic variants included in our analysis were AhR (G1661A, rs2066853), CYP1A2*1K (C163A, rs762551), CYP1B1*3 (G4329C, rs1056836), CYP1B1*4 (A4393G, rs1800440), CYP2C9*2 (C3608T, rs28371674), CYP2C9*3 (A1075C, rs1057910), CYP2E1 (C1055T, rs2031920), SULT1A1*2 (G638A, rs9282861), microsomal epoxide hydrolase (EPHX1, T337C, rs1051740), UGT1A7*9 (G343A, rs61261057), GSTM1 homozygous deletion, GSTT1 homozygous deletion, and 7 NAT2 SNPs: (C282T, rs1041983), (A803G, rs1208), (C481T, rs1799929), (G590A, rs1799930), (G857A, rs1799931), (G191A, rs1801279) and (T341C, rs1801280).
Table I.
Polymorphisms in selected polycyclic aromatic amine metabolizing genes and their effect on enzyme function
| Variant (alleles)a,b | Risk allele frequencyc | Amino acid change | Effect of risk allele on metabolizing enzymatic activity | |
|---|---|---|---|---|
| Phase I enzymes | ||||
| AHR | rs2066853(G/A) | 0.183 | R554K | Higher inducibility (12) |
| CYP1A2 | rs762551 (C/A) | 0.716 | NA | Increasing enzymatic activity (45) |
| CYP1B1 | rs1056836 (G/C) | 0.532 | V432L | Increasing mRNA expression (46) |
| rs1800440 (A/G) | 0.827 | N453S | Increasing enzymatic activity (47) | |
| CYP2C9 | rs28371674 (C/T) | 0.118 | R144C | Increasing enzymatic activity (48)d |
| rs1057910 (A/C) | 0.059 | I359L | Increasing enzymatic activity (49)d | |
| CYP2E1 | rs2031920 (C/T) | 0.022 | NA | Increasing enzymatic activity by enhancing the transcription (50,51) |
| Phase II enzymes | ||||
| NAT1 e | rs4986782 (G/A) | 0.017 | R187Q | Slow acetylation; Decreasing enzymatic activity (52) |
| rs5030839 (C/T) | 0.003 | R187X | Slow acetylation; Decreasing enzymatic activity (52) | |
| NAT2 e | rs1041983 (C/T) | 0.335 | Y94Y | The combined effect of all 7 SNPs were categorized to slow acetylation(reference), intermediate acetylation and rapid acetylation (Increasing enzymatic activity) (52,53) |
| rs1208 (G/A) | 0.426 | R268K | ||
| rs1799929 (C/T) | 0.409 | L161L | ||
| rs1799930 (G/A) | 0.711 | R197Q | ||
| rs1799931 (G/A) | 0.048 | G286E | ||
| rs1801279 (G/A) | 0.008 | R64Q | ||
| rs1801280 (T/C) | 0.570 | I114T | ||
| EPHX1 | rs1051740 (T/C) | 0.290 | T213H | Decreasing enzymatic activity (54) |
| SULT1A1 | rs9282861 (G/A) | 0.236 | R213H | Decreasing enzymatic activity (55,56) |
| UGT1A7 | rs61261057 (G/A) | 0.002 | G115S | Decreasing enzymatic activity (57) |
| GSTM1 | Homozygous deletion | 0.089 | Null | Depleted enzymatic activity (58,59) |
| GSTT1 | Homozygous deletion | 0.145 | Null | Depleted enzymatic activity (59) |
aDatabase of SNPs (http://www.ncbi.nlm.nih.gov/SNP/).
bRisk allele is highlighted in bold.
cFrequency of the minor allele among controls in the TCPS.
dDatabase of SNPs (http://www.cypalleles.ki.se/cyp2c9.htm).
eDatabase of SNPs (http://louisville.edu/medschool/pharmacology/nat/). Details for phenotype imputation shown in Supplementary Appendices 3 and 4, available at Carcinogenesis Online.
We utilized genomic DNA extracted from blood or buccal cells for genotyping assays (15). All allelic gene polymorphisms were assessed by TaqMan OpenArray system. TaqMan OpenArray Assay-on-Demand reagents were available from Applied Biosystems (ABI) for all SNPs except NAT2 G191A (rs1801279). Primers for NAT2 G191A (rs1801279) polymorphisms are self-designed and synthesized by ABI (primers were GGAGTTGGGCTTAGAGGCTATTTT and CAGAAGTTGATTGACCTGGAGACA; probes were VIC-CCACCCCGGTTTC and FAM-CCCACCCTGGTTTC). Primers and probes for these SNPs were pre-loaded by ABI. DNA samples (2.5 μl) and TaqMan Universal PCR Master Mix (2.5 μl) were mixed in a 384-well plate. PCR was performed, consisting of an initial denaturation step at 93°C for 10min and 50 cycles of 95°C for 45 s, 94°C for 13 s and 53°C for 134 s, and post-PCR hold at 25°C for 2min. The fluorescence imaging of genotyping plates could be performed together with the ABI OpenArrayTM NT Imager. Allele frequencies were determined by ABI OpenArray software (AutoCallerTM). Laboratory staffs were blinded to the case–control status of samples. Quality-control protocols for genotyping assays were employed as described previously (15). Briefly, each 384-well plate contained 4 water blanks, 8 CEPH 1347-02 DNA and 16 blinded quality-control samples. Blinded quality-control samples were taken from the second tube of samples included in the study. Quality-control samples were distributed across the 384-well plates. Concordance rate for blinded quality-control samples was 100% for all SNPs. In addition, DNA of 45 Caucasian samples included in the HapMap and Perlegen projects was purchased from Coriell Cell Repositories (http://locus.umdnj.edu/ccr/) and genotyped for all 21 SNPs. Average consistency rates of the 21 SNPs were 99.3% compared with data from HapMap (http://www.hapmap.org) and Perlegen (http://genome.perlegen.com). Call rates for all SNPs were >95% (mean call rate = 97.5 %) and all genotypes were consistent with Hardy–Weinberg equilibrium with P > 0.05 in controls.
Statistical analysis
Each polymorphism was tested in controls to ensure fitting with Hardy–Weinberg equilibrium. We assigned a risk allele to each genetic variant based on the effect on tobacco-carcinogen metabolism. A risk allele was defined as an allele that either increases the activity/amount of a Phase I enzyme or reduces the activity/amount of a Phase II enzyme. For genetic variants other than in NAT1 and NAT2, each participant received a score of 0, 1 or 2 for carrying zero, one or two risk alleles, respectively (Supplementary Appendix 1, available at Carcinogenesis Online). NAT1 and NAT2 phenotype statuses were derived based on the combination of several SNPs provided in Supplementary Appendices 3 and 4, available at Carcinogenesis Online. Predicted NAT1 phenotypes were classified ‘slow’ with a genetic risk score of ‘0’ or ‘non-slow’ with a score of ‘1’. Derived NAT2 phenotypes were classified ‘slow’, ‘intermediate’ and ‘fast’ with a score of ‘0’, ‘1’ and ‘2’, respectively. The tobacco-carcinogen-metabolizing genetic score was derived by summing individual risk scores across all Phase I activation enzymes (AhR, CYP1A2, CYP1B1, CYP2C9 and CYP2E1) and Phase II enzymes (NAT1, NAT2, EPHX1, SULT1A1, UGT1A7, GSTM1 and GSTT1; total possible score = 19). We used tertile cut-points for overall carcinogen-metabolizing scores of ‘9’ and ‘12’ to classify participants into low-risk genetic risk group (score ≤ 9), intermediate genetic risk group (score 10–12) and high-risk genetic risk group (score > 12; Supplementary Appendix 2, available at Carcinogenesis Online).
General linear models and Mantel-Haenszel chi-square tests were used to compare distribution of demographic characteristics and known colorectal-cancer risk factors across carcinogen-metabolizing risk groups, as well as between case and control groups, with adjustment for age and sex when appropriate. Unconditional logistic regression models were used to estimate odds ratios (OR) and their 95% confidence intervals (CI) for the association between genetic variants and smoking with polyp risk. ORs were adjusted for selected confounders for colorectal adenoma that showed significantly different distributions between cases and controls. Variables selected for multivariate analysis included age (40–49, 50–59, 60–64, ≥65), sex, study site (academic medical center, VA medical center), educational attainment (high school or less, some college, college graduate, graduate or professional education), red meat intake (g/day, continuous), regular alcohol consumption (never/former/current), body mass index (BMI, continuous), regular exercise (yes/no), regular NSAID use (never/former/current), year of recruitment and recruitment before or after colonoscopy. P values for linear trend tests were derived by treating categorical variables as continuous parameters in the models (16). Likelihood ratio tests of multiplicative interaction for categorical variables were used to compare models with and without interaction terms (16). P values of ≤0.05 (two-sided probability) were considered statistically significant. All analyses were conducted using SAS statistical software (version 9.2; SAS Institute, Cary, NC).
Results
Table I summarizes Phase I and Phase II tobacco-carcinogen-metabolizing enzymes and functional genetic variants of these enzymes included in this analysis. Highlighted in bold are risk alleles which may increase internal carcinogen exposure, based on results of in vitro functional studies.
Distributions of selected demographic characteristics and major risk factors for colorectal cancer are presented in Table II for the three polyp case groups and the polyp-free controls. More controls than cases were recruited from Vanderbilt Medical Center than the VA medical center. Compared with controls, polyp cases were more likely to be male, smokers and regular alcohol consumers and less probable to use NSAIDs regularly. Cases also had higher BMI and lower educational attainment than controls. Cases with any adenoma were older than controls, whereas cases with HPP only were similar in age to controls. Case–control distributions of race and indication for colonoscopy were comparable.
Table II.
Selected demographic characteristics and major known risk factors for colorectal cancer by study groups, TCPS, 2003–2010
| Characteristic | Controls | Polyp cases | |||
|---|---|---|---|---|---|
| (N = 3336) | Adenoma only (N = 1160) | HPP only (N = 532) | Both (N = 368) | P valuea | |
| Study site (%) | |||||
| Vanderbilt University | 73.8 | 61.3 | 54.0 | 47.3 | <0.001 |
| Veterans Affairs | 26.2 | 38.7 | 46.0 | 52.7 | |
| Age [years, mean (SD)] | 57.2 (7.6) | 59.2 (7.4) | 57.1 (6.9) | 58.9 (6.5) | <0.001 |
| Sex (female, %) | 43.8 | 27.2 | 30.1 | 18.8 | <0.001 |
| Indications for colonoscopy (%)b | |||||
| Screening | 58.6 | 57.7 | 56.1 | 52.4 | 0.062 |
| Other | 41.4 | 42.3 | 43.9 | 47.6 | |
| Educational attainment (%)b | |||||
| High school or less | 24.5 | 29.7 | 33.1 | 37.7 | <0.001 |
| Some college | 28.7 | 27.0 | 31.4 | 38.5 | |
| College graduate | 20.4 | 21.9 | 18.5 | 12.7 | |
| Graduate or professional education | 26.5 | 21.4 | 17.0 | 11.0 | |
| Race (white, %) | 89.6 | 88.7 | 91.9 | 91.9 | 0.114 |
| Colorectal cancer family history (%)b | 8.4 | 9.9 | 8.1 | 10.4 | 0.378 |
| Regular cigarette smoking (%)b | 49.1 | 57.0 | 72.6 | 75.7 | <0.001 |
| Regular alcohol consumption (%)b | 43.1 | 47.0 | 50.7 | 45.2 | 0.009 |
| Body mass index (kg/m2, mean) | 28.1 | 28.7 | 28.8 | 29.1 | <0.001 |
| Regularly exercised (%)b | 57.5 | 52.8 | 53.3 | 48.0 | <0.001 |
| Regular NSAID use (%)b | 51.6 | 52.8 | 53.3 | 48.0 | 0.025 |
| Red meat intake (g/day, mean) | 57.4 | 73.4 | 79.8 | 85.3 | <0.001 |
Both, synchronous adenoma and HPPs.
aDerived from analysis of variance for continuous variables and chi-square test for categorical variables.
bStandardized by age (40–49, 50–59, 60–64 and ≥65 years) and sex distribution of all study participants.
No apparent association was observed between the tobacco-carcinogen-metabolizing risk score and the risk of any-polyp groups evaluated in the study including adenoma only, HPP only and those with both synchronous adenoma and HPP (Table III). Associations between risk of polyps and the Phase I or Phase II carcinogen-metabolizing risk score were also null (data not shown).
Table III.
Association between carcinogen-metabolizing scores and colorectal polyp risk, the TCPSa
| Carcinogen-metabolizing risk score by tertile | Control (N = 3,336) | Polyp cases | |||||||
|---|---|---|---|---|---|---|---|---|---|
| All polyps (N = 2060) | Adenoma only (N = 1160) | HPP only (N = 532) | Both (N = 368) | ||||||
| n | OR (95%CI)a | n | OR (95%CI)a | n | OR (95%CI)a | n | OR (95%CI)1 | ||
| T1 (0–9) | 1218 | 756 | 1.0 (reference) | 427 | 1.0 (reference) | 186 | 1.0 (reference) | 143 | 1.0 (reference) |
| T2(10–12) | 1775 | 1075 | 1.0 (0.9–1.1) | 611 | 1.0 (0.8–1.1) | 288 | 1.1 (0.9–1.3) | 180 | 0.8 (0.7–1.1) |
| T3(13–19) | 343 | 225 | 1.1 (0.9–1.3) | 122 | 1.1 (0.8–1.3) | 58 | 1.2 (0.9–1.6) | 45 | 1.3 (0.9–1.9) |
| Trend test | P = 0.654 | P = 0.933 | P = 0.287 | P = 0.096 | |||||
Both, synchronous adenoma and HPP.
aAdjusted for age, sex, study sites, educational attainment, alcohol consumption, BMI, physical activity, regular NSAID use, red meat intake, total energy intake, year of recruitment and recruitment before or after colonoscopy.
All cigarette-smoking variables were associated with risk of all polyps combined regardless of the category of carcinogen-metabolizing risk score (Table IV). Association with cigarette smoking was stronger among participants with a high carcinogen-metabolizing risk than those with a low score. Tests for multiplicative interactions were statistically significant for pack-years of cigarette smoking (P interaction = 0.032) and cigarettes smoked per day (P interaction = 0.025). No apparent interaction between genetic risk score and years of smoking was observed and thus it is probable that the significant interaction with pack-years smoked may be explained primarily by intensity but not duration of cigarette smoking. Because of this, only results for smoking status and number of cigarettes smoked per day are presented in subsequent tables.
Table IV.
Risk of all polyps combined in association with cigarette smoking by carcinogen-metabolizing risk category, the TCPS, 2003–2010
| Smoking | Carcinogen-metabolizing risk score | |||||
|---|---|---|---|---|---|---|
| Low risk | Intermediate risk | High risk | ||||
| n a | OR (95% CI)b | n a | OR (95% CI)b | n a | OR (95% CI)b | |
| Smoking status | ||||||
| Never | 261/656 | 1.0 (reference) | 344/916 | 0.9 (0.8–1.1) | 67/182 | 0.9 (0.7–1.3) |
| Former | 286/405 | 1.4 (1.1–1.7) | 406/630 | 1.3 (1.1–1.6) | 88/117 | 1.6 (1.2–2.2) |
| Current | 207/156 | 2.9 (2.2–3.8) | 328/227 | 3.0 (2.4–3.9) | 69/43 | 3.4 (2.2–5.2) |
| Test of interaction, P = 0.341 | ||||||
| Pack-years | ||||||
| Never | 261/656 | 1.0 (reference) | 344/916 | 0.9 (0.8–1.1) | 67/182 | 0.9 (0.7–1.3) |
| 1–9 | 101/156 | 1.6 (1.2–2.1) | 137/277 | 1.2 (0.9–1.6) | 36/54 | 1.6 (1.0–2.5) |
| 10–29 | 149/178 | 1.8 (1.4–2.4) | 213/268 | 1.7 (1.3–2.1) | 34/65 | 1.2 (0.7–1.8) |
| ≥30 | 242/227 | 2.0 (1.5–2.5) | 380/307 | 2.3 (1.8–2.9) | 84/40 | 4.0 (2.6–6.1) |
| Test of interaction, P = 0.032 | ||||||
| Cigarettes per day | ||||||
| Never | 261/656 | 1.0 (reference) | 344/916 | 0.9 (0.8–1.1) | 67/182 | 1.0 (0.7–1.3) |
| 1–19 | 183/211 | 2.1 (1.6–2.7) | 269/345 | 1.8 (1.4–2.2) | 44/75 | 1.4 (0.9–2.1) |
| 20–29 | 151/187 | 1.5 (1.1–1.9) | 233/292 | 1.5 (1.2–1.9) | 64/54 | 2.3 (1.6–3.5) |
| ≥30 | 159/163 | 1.7 (1.3–2.2) | 230/215 | 1.8 (1.4–2.3) | 49/30 | 2.9 (1.8–4.8) |
| Test of interaction, P = 0.025 | ||||||
| Years of cigarette smoking | ||||||
| Never | 261/656 | 1.0 (reference) | 344/916 | 0.9 (0.8–1.1) | 67/182 | 1.0 (0.7–1.3) |
| 1–14 | 86/159 | 1.2 (0.9–1.6) | 110/284 | 0.9 (0.7–1.2) | 35/54 | 1.6 (1.0–2.6) |
| 15–24 | 83/122 | 1.4 (1.0–2.0) | 104/158 | 1.3 (1.0–1.8) | 21/32 | 1.5 (0.8–2.6) |
| 25–34 | 111/121 | 2.0 (1.5–2.7) | 181/185 | 2.1 (1.6–2.8) | 35/35 | 2.1 (1.3–3.6) |
| ≥35 | 212/159 | 2.4 (1.8–3.3) | 337/229 | 2.9 (2.3–3.7) | 63/39 | 3.2 (2.2–4.7) |
| Test of interaction, P = 0.294 | ||||||
an, number of cases/controls. The total number of subjects was 5380 (including 2054 cases and 3326 controls). Sixteen subjects were deleted because of missing data.
bAdjusted for age, sex, study sites, educational attainment, alcohol consumption, BMI, physical activity, regular NSAID use, red meat intake, total energy intake, year of recruitment and recruitment before or after colonoscopy.
Associations with cigarette smoking by carcinogen-metabolizing risk categories were analyzed separately for three case groups: (i) adenoma only, (ii) HPP only and (iii) both synchronous adenoma and HPP (Table V). Multiplicative interactions at P ≤ 0.05 were found for number of cigarettes smoked per day in the adenoma-only and HPP-only groups. Although associations with smoking were stronger for those with both synchronous adenoma and HPP than with the adenoma-only or HPP-only groups, no significant interaction was observed for both synchronous adenoma and HPP. Similar to the results for all polyps combined, no apparent interaction was observed for years of smoking in these subgroup analyses (data not shown).
Table V.
Joint association of cigarette smoking with risk of polyps by subtypes, the TCPS, 2003–2010
| Cigarette smoking | Carcinogen-metabolizing risk category | |||||
|---|---|---|---|---|---|---|
| Low risk | Intermediate risk | High risk | ||||
| n a | OR (95% CI)b | n a | OR (95% CI)b | n a | OR (95% CI)b | |
| Adenoma only (N = 1161) | ||||||
| Smoking status | ||||||
| Never | 183/656 | 1.0 (reference) | 237/916 | 0.9 (0.7–1.1) | 43/182 | 0.9 (0.6–1.3) |
| Former | 162/405 | 1.1 (0.8–1.4) | 230/630 | 1.0 (0.8–1.3) | 48/117 | 1.2 (0.8–1.8) |
| Current | 81/156 | 1.6 (1.1–2.3) | 144/227 | 1.9 (1.4–2.6) | 30/43 | 2.1 (1.2–3.5) |
| Test of interaction, P = 0.141 | ||||||
| Cigarettes per day | ||||||
| Never | 183/656 | 1.0 (reference) | 237/916 | 0.9 (0.7–1.1) | 43/182 | 0.9 (0.6–1.3) |
| 1–19 | 92/211 | 1.4 (1.0–1.9) | 140/345 | 1.3 (1.0–1.7) | 25/75 | 1.1 (0.6–1.8) |
| 20–29 | 74/187 | 1.1 (0.8–1.5) | 110/292 | 1.0 (0.8–1.4) | 27/54 | 1.5 (0.9–2.4) |
| ≥30 | 77/163 | 1.1 (0.8–1.6) | 123/215 | 1.4 (1.0–1.9) | 26/30 | 2.0 (1.1–3.5) |
| Test of interaction, P = 0.039 | ||||||
| HPP only (N = 532) | ||||||
| Smoking status | ||||||
| Never | 50/656 | 1.0 (reference) | 74/916 | 1.1 (0.7–1.6) | 13/182 | 1.0 (0.5–1.8) |
| Ever | 71/405 | 2.0 (1.3–2.9) | 110/630 | 2.0 (1.4–2.9) | 23/117 | 2.4 (1.4–4.2) |
| Current | 64/156 | 5.1 (3.3–7.8) | 103/227 | 5.4 (3.6–8.0) | 22/43 | 6.1 (3.3–11.2) |
| Test of interaction, P = 0.753 | ||||||
| Cigarettes per day | ||||||
| Never | 50/656 | 1.0 (reference) | 74/916 | 1.1 (0.7–1.6) | 13/182 | 1.0 (0.5–1.8) |
| 1–19 | 63/211 | 3.9 (2.6–5.9) | 83/345 | 3.0 (2.1–4.4) | 12/75 | 1.9 (0.9–3.8) |
| 20–29 | 36/187 | 1.9 (1.2–3.0) | 73/292 | 2.7 (1.8–4.0) | 17/54 | 3.3 (1.8–6.3) |
| ≥30 | 36/163 | 2.1 (1.3–3.4) | 56/215 | 2.4 (1.5–3.7) | 16/30 | 6.1 (3.1–12.2) |
| Test of interaction, P = 0.024 | ||||||
| Both (N = 368) | ||||||
| Smoking status | ||||||
| Never | 28/656 | 1.0 (reference) | 33/916 | 0.8 (0.5–1.4) | 11/182 | 1.5 (0.7–3.2) |
| Ever | 53/405 | 2.1 (1.3–3.5) | 66/630 | 1.7 (1.1–2.8) | 17/117 | 2.9 (1.5–5.6) |
| Current | 62/156 | 7.7 (4.6–12.9) | 81/227 | 6.2 (3.8–10.1) | 17/43 | 7.4 (3.6–15.1) |
| Test of interaction, P = 0.410 | ||||||
| Cigarettes per day | ||||||
| Never | 28/656 | 1.0 (reference) | 33/916 | 0.8 (0.5–1.4) | 11/182 | 1.5 (0.7–3.2) |
| 1–19 | 28/211 | 2.9 (1.6–5.0) | 46/345 | 2.7 (1.6–4.5) | 7/75 | 2.3 (0.9–5.5) |
| 20–29 | 41/187 | 3.5 (2.0–5.9) | 50/292 | 2.9 (1.7–4.7) | 20/54 | 6.0 (3.1–11.7) |
| ≥30 | 46/163 | 3.8 (2.2–6.4) | 51/215 | 3.1 (1.8–5.2) | 7/30 | 3.9 (1.5–9.9) |
| Test of interaction, P = 0.837 | ||||||
Both, synchronous adenoma and HPP.
an, number of cases/controls. Count may not sum up to total because of missing data.
bAdjusted for age, sex, study sites, educational attainment, alcohol consumption, BMI, physical activity, regular NSAID use, red meat intake, total energy intake, year of recruitment, and recruitment before or after colonoscopy.
Strong association and significant interaction of carcinogen-metabolizing genetic risk score and smoking were found for clinically important adenomas, that is advanced or multiple adenomas, but not for non-advanced and single adenomas (Table VI). Most of the joint effects for non-advanced and single adenomas were not significant, although the sample size (N = 578) was slightly larger than advanced or multiple adenomas group (N = 531).
Table VI.
Joint association of cigarette smoking and carcinogen-metabolizing risk categories in relation to adenoma risk, the TCPS, 2003–2010
| Cigarette smoking | Total metabolizing risk category | |||||
|---|---|---|---|---|---|---|
| Low risk | Intermediate risk | High risk | ||||
| n a | OR (95% CI)b | n a | OR (95% CI)b | n a | OR (95% CI)b | |
| Advanced or multiple adenomas (N = 531) | ||||||
| Smoking status | ||||||
| Never | 117/656 | 1.0 (reference) | 131/916 | 1.1 (0.8–1.5) | 27/182 | 0.9 (0.5–1.6) |
| Ever | 89/405 | 1.3 (0.9–1.9) | 117/630 | 1.3 (0.9–1.9) | 27/117 | 1.4 (0.8–2.4) |
| Current | 46/155 | 2.0 (1.2–3.2) | 63/227 | 3.1 (2.1–4.6) | 11/43 | 3.7 (2.0–7.0) |
| Test of interaction, P = 0.010 | ||||||
| Cigarettes per day | ||||||
| Never | 117/656 | 1.0 (reference) | 131/916 | 1.1 (0.8–1.5) | 27/182 | 0.9 (0.5–1.6) |
| 1–19 | 56/210 | 1.6 (1.0–2.4) | 73/345 | 1.8 (1.2–2.6) | 12/75 | 1.5 (0.7–2.9) |
| 20–29 | 41/187 | 1.3 (0.8–2.1) | 56/292 | 1.3 (0.9–2.0) | 11/54 | 2.4 (1.3–4.5) |
| ≥30 | 38/163 | 1.5 (0.9–2.4) | 51/215 | 2.2 (1.4–3.2) | 15/30 | 2.3 (1.1–4.9) |
| Test of interaction, P = 0.132 | ||||||
| Single non-advanced adenomas (N = 578) | ||||||
| Smoking status | ||||||
| Never | 111/656 | 1.0 (reference) | 123/916 | 0.8 (0.6–1.1) | 26/182 | 0.9 (0.6–1.5) |
| Ever | 84/405 | 1.0 (0.7–1.4) | 108/630 | 0.8 (0.6–1.1) | 23/117 | 1.0 (0.6–1.6) |
| Current | 40/156 | 1.3 (0.9–2.1) | 52/227 | 1.1 (0.7–1.6) | 10/43 | 1.1 (0.5–2.4) |
| Test of interaction, P = 0.791 | ||||||
| Cigarettes per day | ||||||
| Never | 111/656 | 1.0 (reference) | 123/916 | 0.8 (0.6–1.1) | 26/182 | 0.9 (0.6–1.5) |
| 1–19 | 52/211 | 1.4 (0.9–2.0) | 65/345 | 1.0 (0.7–1.4) | 11/75 | 0.8 (0.4–1.6) |
| 20–29 | 39/187 | 0.9 (0.6–1.4) | 52/292 | 0.8 (0.6–1.2) | 11/54 | 1.0 (0.5–2.0) |
| ≥30 | 33/163 | 0.9 (0.6–1.4) | 43/215 | 0.8 (0.6–1.3) | 11/30 | 1.5 (0.7–3.1) |
| Test of interaction, P = 0.159 | ||||||
an, number of cases/controls. The total number of subjects was 4435 (including 1109 cases with adenoma only and 3326 controls). Count may not sum up to total because of missing data.
bAdjusted for age, sex, study sites, educational attainment, alcohol consumption, BMI, physical activity, regular NSAID use, red meat intake, total energy intake, year of recruitment, and recruitment before or after colonoscopy.
Discussion
In this study, we showed that genetically controlled tobacco-carcinogen-metabolizing pattern modifies the association between cigarette smoking and the risk of colorectal polyps, including adenomas and HPP. Among patients with adenomas only, this interaction was found for high-risk adenomas, that is advanced or multiple adenomas, but not for low-risk adenomas, that is non-advanced and single adenomas. Our results provide strong evidence to implicate an etiological role of tobacco-carcinogen exposure in the formation and in the progression of colorectal neoplasia.
We did not find any association of polyp risk with any individual genetic variant or with combined risk scores of these variants. Results from previous studies regarding the association of genetic factors with colorectal cancer risk (17–23) and adenoma risk (24–36) were largely null. Results for potential modifying effects of genetic factors on the association of cigarette smoking have been inconsistent. These inconsistent findings are not surprising because each of the genetic variants in the metabolizing pathway plays a small role in the activation or detoxification of tobacco carcinogens. Therefore, it is important to combine information from multiple genetic variants to capture the carcinogen-metabolizing pattern of each individual. In addition, because multiple comparisons have been made and many previous studies are small in sample size, the inconsistent results are hard to interpret. Therefore, in this study, we did not attempt to address interactions between individual genetic variants and cigarette smoking. No association between polyp risk and the genetic score was observed. However, a significant effect modification of the genetic score on the association between cigarette smoking and risk of colorectal polyps was found. We found that the effect modification of genetic factors on cigarette smoking may be due more to smoking intensity than smoking duration. Although this observation is not unexpected because internal exposure dose is more probable to be affected by intensity than duration of exposure, future research is needed to confirm this finding.
Our study, in general, fulfilled the three criteria required to apply Mendelian randomization analyses (39–42). First, the genotypic profile, as summarized by the carcinogen-metabolizing risk score, was independent of factors that may confound the association between cigarette-smoking intake levels or carcinogen-exposure levels and risk of colorectal polyps (39–41). We evaluated this criterion by analyzing the distribution of demographic characteristics and risk factors presented in Table I by carcinogen-metabolizing risk categories (Supplementary Appendix 2, available at Carcinogenesis Online). Our results show no difference between the low-, intermediate- and high-carcinogen-metabolizing risk categories in the distributions of all demographic characteristics and risk factors. These findings support successful Mendelian randomization of study participants. Regarding the second criterion, the carcinogen genetic risk profile must be related to risk of disease through the exposure of interest only and not through other pathways. Because many metabolizing genes are pleiotropic, they may be involved in the metabolism of other carcinogens. We selected functional SNPs involved in the metabolism of multiple tobacco carcinogens, including aromatic amines, polycyclic aromatic hydrocarbons and nitrosamines. In an attempt to measure the overall genetic variation in the metabolism of multiple tobacco carcinogens, we combined these SNPs to construct a genetic score. We believe this approach is appropriate because the objective of this study was to evaluate interaction with tobacco smoking, not interaction with any single tobacco carcinogen. Furthermore, because multiple tobacco carcinogens have been linked to the risk of colorectal polyps, it would be more powerful to evaluate the overall association of polyp risk with tobacco smoke than the specific association of polyp risk with any single tobacco carcinogen. On the other hand, because these SNPs were selected from the pathway of tobacco-carcinogen metabolism, it is highly unlikely that the genetic score derived from the entire set of tobacco-carcinogen-metabolizing SNPs would be closely related to the metabolism of substrates other than tobacco carcinogens. In other words, the carcinogen-metabolizing risk score constructed in this study should be more specific for tobacco-carcinogen metabolism than any single variant of carcinogen-metabolizing enzymes. We also evaluated the interaction of carcinogen-metabolizing risk score with red meat intake, NSAID use, alcohol intake, BMI and physical activity, and found no interactions between these factors and polyp risk. Therefore, it is unlikely that potential pleiotropic effects of certain genes could appreciably affect the results from this study. Finally, based on the knowledge of carcinogen-metabolizing patterns, the carcinogen-metabolizing risk score derived from this study should affect the internal effective dose of carcinogen exposure and thus can further define internal-exposure status for people with the same external exposure level of tobacco carcinogen. Interestingly, the association of cancer risk with cigarette smoking (or aromatic amines) and carcinogen-metabolizing enzymes is among the examples provided by Smith (43) and Thomas (44) for Mendelian randomization analyses of gene–environment interactions.
We assigned a carcinogen-metabolizing score of ‘0’, ‘1’ or ‘2’ for genotypes carrying zero, one or two risk alleles, respectively. Ideally, this score should be constructed by taking into consideration the effect size of each isozyme in the pathway of tobacco-carcinogen metabolism. However, because in vitro experiments to assess the function and effect size of these isozymes were conducted under various conditions, it would be difficult to use these data to construct a genetic score. To avoid over-fitting the data, we did not use the OR for each genetic variant as the weight to construct the genetic score. However, this simple approach allowed us to identify significant interactions. If we could take into consideration the effect size of each isozyme in construction of the carcinogen-metabolizing score, we expect that a stronger interaction between genetic factors and cigarette smoking might be observed.
This study focused on precursors for colorectal cancer, that is colorectal polyps, thus eliminating possible survival bias commonly encountered in cancer case–control studies as a result of failure to recruit patients with short survival time. Because these polyps were benign lesions, recall bias and any subsequent lifestyle change following polyp diagnosis are unlikely to be substantial. Other important strengths of this study are the use of colonoscopy to define patient groups and large sample size. The majority of study participants (87.8%) were recruited prior to colonoscopy and thus prior to polyp diagnosis, which reduces possible selection bias. Exclusion of participants recruited after colonoscopy (n = 894) did not appreciably change the associations observed. Response rates are not optimal in this study, which could introduce selection bias. However, as discussed previously, these potential biases can be reduced in our study through the use of Mendelian randomization analysis. Because we did not have data regarding internal exposure of tobacco carcinogens, the internal exposure level in our study was inferred using data from exposure assessment and the carcinogen-metabolizing score.
In summary, our findings suggest that genetic variants in carcinogen-metabolizing enzymes may modify the association of cigarette smoking with colorectal polyp risk. We expect additional functional variants in the carcinogen-metabolizing pathway will be identified in the future to improve the classification of participants into low- to high-carcinogen-metabolizing risk categories. In other words, the true synergistic effect of genetic factors and cigarette smoking could be stronger than what we observed in this study.
Supplementary material
Supplementary Appendices 1–4 can be found at http://carcin.oxfordjournals.org/
Funding
National Cancer Institute (P50CA950103 and R01CA97386); Biospe cimen and Survey Research Shared Resource (P30CA68485).
Conflicts of Interest Statement: None declared.
Supplementary Material
Acknowledgement
We would like to thank the study participants and research staff of the Tennessee Colorectal Polyp Study for their support of this research.
The authors’ responsibilities were as follows: W.Z. conceived and designed the study; Z.F. analyzed the data; Z.F. and W.Z. drafted the manuscript; M.J.S. coordinated the field operation of the parent study and contributed to study design and manuscript revision; G.L. performed the genotyping; R.M.N. and W.E.S. provided support for the clinical operation; D.H. provided support to derive NAT1 and NAT2 phenotypes; and all authors reviewed and approved the final draft. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Glossary
Abbreviations:
- CI
confidence intervals
- CYP
cytochrome P450 enzymes
- EPHX
microsomal epoxide hydrolases
- GST
glutathione-S-transferases
- HPP
hyperplastic polyp
- NAT
N-acetyltransferases
- OR
odds ratio
- SNP
single nucleotide polymorphisms
- TCPS
Tennessee Colorectal Polyp Study
- UGT
UDP-glucuronosyltransferases.
References
- 1. Botteri E, et al. (2008). Smoking and colorectal cancer: a meta-analysis. JAMA, 300, 2765–2778 [DOI] [PubMed] [Google Scholar]
- 2. Edoardo B, et al. (2008). Cigarette smoking and adenomatous polyps: a meta-analysis. Gastroenterology, 134, 388–395.e3 [DOI] [PubMed] [Google Scholar]
- 3. Muto T, et al. (1975). The evolution of cancer of the colon and rectum. Cancer, 36, 2251–2270 [DOI] [PubMed] [Google Scholar]
- 4. Stryker S.J, et al. (1987). Natural history of untreated colonic polyps. Gastroenterology, 93, 1009–1013 [DOI] [PubMed] [Google Scholar]
- 5. Fearon E.R, et al. (1990). A genetic model for colorectal tumorigenesis. Cell, 61, 759–767 [DOI] [PubMed] [Google Scholar]
- 6. Hawkins N.J, et al. (2001). Sporadic colorectal cancers with microsatellite instability and their possible origin in hyperplastic polyps and serrated adenomas. J. Natl. Cancer Inst., 93, 1307–1313 [DOI] [PubMed] [Google Scholar]
- 7. Leggett B, et al. (2010). Role of the serrated pathway in colorectal cancer pathogenesis. Gastroenterology, 138, 2088–2100 [DOI] [PubMed] [Google Scholar]
- 8. Chu D.Z, et al. (1986). The significance of synchronous carcinoma and polyps in the colon and rectum. Cancer, 57, 445–450 [DOI] [PubMed] [Google Scholar]
- 9. Kellokumpu I, et al. (1991). Multiple adenomas and synchronous hyperplastic polyps as predictors of metachronous colorectal adenomas. Ann. Chir. Gynaecol., 80, 30–35 [PubMed] [Google Scholar]
- 10. Hecht S.S. (2003). Tobacco carcinogens, their biomarkers and tobacco-induced cancer. Nat. Rev. Cancer, 3, 733–744 [DOI] [PubMed] [Google Scholar]
- 11. Nebert D.W. (1991). Role of genetics and drug metabolism in human cancer risk. Mutat. Res., 247, 267–281 [DOI] [PubMed] [Google Scholar]
- 12. Fujii-Kuriyama Y, et al. (2005). Molecular mechanisms of AhR functions in the regulation of cytochrome P450 genes. Biochem. Biophys. Res. Commun., 338, 311–317 [DOI] [PubMed] [Google Scholar]
- 13. Shrubsole M.J, et al. (2008). Alcohol drinking, cigarette smoking, and risk of colorectal adenomatous and hyperplastic polyps. Am. J. Epidemiol., 167, 1050–1058 [DOI] [PubMed] [Google Scholar]
- 14. Fu Z, et al. (2011). Association of meat intake and meat-derived mutagen exposure with the risk of colorectal polyps by histologic type. Cancer Prev. Res. (Phila)., 4, 1686–1697 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Zheng W, et al. (2009). Genome-wide association study identifies a new breast cancer susceptibility locus at 6q25.1. Nat. Genet., 41, 324–328 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Madure M, et al. (1992). Tests for trend and dose response: misinterpretations and alternatives. Am. J. Epidemiol., 135, 96–104 [DOI] [PubMed] [Google Scholar]
- 17. Ulrich C.M, et al. (2001). Epoxide hydrolase Tyr113His polymorphism is associated with elevated risk of colorectal polyps in the presence of smoking and high meat intake. Cancer Epidemiol. Biomarkers Prev., 10, 875–882 [PubMed] [Google Scholar]
- 18. Ishibe N, et al. (2002). Genetic polymorphisms in heterocyclic amine metabolism and risk of colorectal adenomas. Pharmacogenetics, 12, 145–150 [DOI] [PubMed] [Google Scholar]
- 19. Chan A.T, et al. (2004). A prospective study of genetic polymorphisms in the cytochrome P-450 2C9 enzyme and the risk for distal colorectal adenoma. Clin. Gastroenterol. Hepatol., 2, 704–712 [DOI] [PubMed] [Google Scholar]
- 20. Tiemersma E.W, et al. (2004). Risk of colorectal adenomas in relation to meat consumption, meat preparation, and genetic susceptibility in a Dutch population. Cancer Causes Control, 15, 225–236 [DOI] [PubMed] [Google Scholar]
- 21. Goode E.L, et al. (2007). Inherited variation in carcinogen-metabolizing enzymes and risk of colorectal polyps. Carcinogenesis, 28, 328–341 [DOI] [PubMed] [Google Scholar]
- 22. Shin A, et al. (2008). Meat intake, heterocyclic amine exposure, and metabolizing enzyme polymorphisms in relation to colorectal polyp risk. Cancer Epidemiol. Biomarkers Prev., 17, 320–329 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Wang H, et al. (2011). Genetic variation in the bioactivation pathway for polycyclic hydrocarbons and heterocyclic amines in relation to risk of colorectal neoplasia. Carcinogenesis, 32, 203–209 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Welfare M.R, et al. (1997). Relationship between acetylator status, smoking, and diet and colorectal cancer risk in the north-east of England. Carcinogenesis, 18, 1351–1354 [DOI] [PubMed] [Google Scholar]
- 25. Chen J, et al. (1998). A prospective study of N-acetyltransferase genotype, red meat intake, and risk of colorectal cancer. Cancer Res., 58, 3307–3311 [PubMed] [Google Scholar]
- 26. Le Marchand L, et al. (2001). Combined effects of well-done red meat, smoking, and rapid N-acetyltransferase 2 and CYP1A2 phenotypes in increasing colorectal cancer risk. Cancer Epidemiol. Biomarkers Prev., 10, 1259–1266 [PubMed] [Google Scholar]
- 27. Le Marchand L, et al. (2002). Well-done red meat, metabolic phenotypes and colorectal cancer in Hawaii. Mutat. Res., 506-507, 205–214 [DOI] [PubMed] [Google Scholar]
- 28. Tiemersma E.W, et al. (2002). Meat consumption, cigarette smoking, and genetic susceptibility in the etiology of colorectal cancer: results from a Dutch prospective study. Cancer Causes Control, 13, 383–393 [DOI] [PubMed] [Google Scholar]
- 29. Barrett J.H, et al. ; Colorectal Cancer Study Group (2003). Investigation of interaction between N-acetyltransferase 2 and heterocyclic amines as potential risk factors for colorectal cancer. Carcinogenesis, 24, 275–282 [DOI] [PubMed] [Google Scholar]
- 30. Butler L.M, et al. (2005). Joint effects between UDP-glucuronosyltransferase 1A7 genotype and dietary carcinogen exposure on risk of colon cancer. Cancer Epidemiol. Biomarkers Prev., 14, 1626–1632 [DOI] [PubMed] [Google Scholar]
- 31. Chan A.T, et al. (2005). Prospective study of N-acetyltransferase-2 genotypes, meat intake, smoking and risk of colorectal cancer. Int. J. Cancer, 115, 648–652 [DOI] [PubMed] [Google Scholar]
- 32. Lilla C, et al. (2006). Effect of NAT1 and NAT2 genetic polymorphisms on colorectal cancer risk associated with exposure to tobacco smoke and meat consumption. Cancer Epidemiol. Biomarkers Prev., 15, 99–107 [DOI] [PubMed] [Google Scholar]
- 33. Küry S, et al. (2007). Combinations of cytochrome P450 gene polymorphisms enhancing the risk for sporadic colorectal cancer related to red meat consumption. Cancer Epidemiol. Biomarkers Prev., 16, 1460–1467 [DOI] [PubMed] [Google Scholar]
- 34. Butler L.M, et al. (2008). Modification by N-acetyltransferase 1 genotype on the association between dietary heterocyclic amines and colon cancer in a multiethnic study. Mutat. Res., 638, 162–174 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Cotterchio M, et al. (2008). Red meat intake, doneness, polymorphisms in genes that encode carcinogen-metabolizing enzymes, and colorectal cancer risk. Cancer Epidemiol. Biomarkers Prev., 17, 3098–3107 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Nöthlings U, et al. (2009). Meat and heterocyclic amine intake, smoking, NAT1 and NAT2 polymorphisms, and colorectal cancer risk in the multiethnic cohort study. Cancer Epidemiol. Biomarkers Prev., 18, 2098–2106 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Moore L.E, et al. (2011). GSTM1 null and NAT2 slow acetylation genotypes, smoking intensity and bladder cancer risk: results from the New England bladder cancer study and NAT2 meta-analysis. Carcinogenesis, 32, 182–189 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Le Marchand L. (2005). The predominance of the environment over genes in cancer causation: implications for genetic epidemiology. Cancer Epidemiol. Biomarkers Prev., 14, 1037–1039 [DOI] [PubMed] [Google Scholar]
- 39. Smith G.D, et al. (2004). Mendelian randomization: prospects, potentials, and limitations. Int. J. Epidemiol., 33, 30–42 [DOI] [PubMed] [Google Scholar]
- 40. Tobin M.D, et al. (2004). Commentary: development of Mendelian randomization: from hypothesis test to ‘Mendelian deconfounding’. Int. J. Epidemiol., 33, 26–29 [DOI] [PubMed] [Google Scholar]
- 41. Thanassoulis G, et al. (2009). Mendelian randomization: nature’s randomized trial in the post-genome era. JAMA, 301, 2386–2388 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Benn M, et al. (2011). Low-density lipoprotein cholesterol and the risk of cancer: a mendelian randomization study. J. Natl. Cancer Inst., 103, 508–519 [DOI] [PubMed] [Google Scholar]
- 43. Smith G.D. Mendelian randomization for strengthening causal inference in observational studies. Perspect. Psychol. Sci., 5, 527–545 [DOI] [PubMed] [Google Scholar]
- 44. Thomas D.C, et al. (2004). Commentary: the concept of ‘Mendelian Randomization’. Int. J. Epidemiol., 33, 21–25 [DOI] [PubMed] [Google Scholar]
- 45. Sachse C, et al. (1999). Functional significance of a C–>A polymorphism in intron 1 of the cytochrome P450 CYP1A2 gene tested with caffeine. Br. J. Clin. Pharmacol., 47, 445–449 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Landi M.T, et al. (2005). CYP1A1 and CYP1B1 genotypes, haplotypes, and TCDD-induced gene expression in subjects from Seveso, Italy. Toxicology, 207, 191–202 [DOI] [PubMed] [Google Scholar]
- 47. Shimada T, et al. (1999). Catalytic properties of polymorphic human cytochrome P450 1B1 variants. Carcinogenesis, 20, 1607–1613 [DOI] [PubMed] [Google Scholar]
- 48. Crespi C.L, et al. (1997). The R144C change in the CYP2C9*2 allele alters interaction of the cytochrome P450 with NADPH:cytochrome P450 oxidoreductase. Pharmacogenetics, 7, 203–210 [DOI] [PubMed] [Google Scholar]
- 49. Takanashi K, et al. (2000). CYP2C9 Ile359 and Leu359 variants: enzyme kinetic study with seven substrates. Pharmacogenetics, 10, 95–104 [DOI] [PubMed] [Google Scholar]
- 50. Hayashi S, et al. (1991). Genetic polymorphisms in the 5’-flanking region change transcriptional regulation of the human cytochrome P450IIE1 gene. J. Biochem., 110, 559–565 [DOI] [PubMed] [Google Scholar]
- 51. Liu Y, et al. (2009). Genetic polymorphism and mRNA levels of cytochrome P450IIE1 and glutathione S-transferase P1 in patients with alcoholic liver disease in different nationalities. HBPD INT, 8, 162–167 [PubMed] [Google Scholar]
- 52. Hein D.W. (2002). Molecular genetics and function of NAT1 and NAT2: role in aromatic amine metabolism and carcinogenesis. Mutat. Res., 506-507, 65–77 [DOI] [PubMed] [Google Scholar]
- 53. Metry K.J, et al. (2010). Effect of rapid human N-acetyltransferase 2 haplotype on DNA damage and mutagenesis induced by 2-amino-3-methylimidazo-[4,5-f]quinoline (IQ) and 2-amino-3,8-dimethylimidazo-[4,5-f]quinoxaline (MeIQx). Mutat. Res., 684, 66–73 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54. Hassett C, et al. (1994). Human microsomal epoxide hydrolase: genetic polymorphism and functional expression in vitro of amino acid variants. Hum. Mol. Genet., 3, 421–428 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Raftogianis R.B, et al. (1997). Phenol sulfotransferase pharmacogenetics in humans: association of common SULT1A1 alleles with TS PST phenotype. Biochem. Biophys. Res. Commun., 239, 298–304 [DOI] [PubMed] [Google Scholar]
- 56. Raftogianis R.B, et al. (1999). Human phenol sulfotransferases SULT1A2 and SULT1A1: genetic polymorphisms, allozyme properties, and human liver genotype-phenotype correlations. Biochem. Pharmacol., 58, 605–616 [DOI] [PubMed] [Google Scholar]
- 57. Villeneuve L, et al. (2003). Novel functional polymorphisms in the UGT1A7 and UGT1A9 glucuronidating enzymes in Caucasian and African-American subjects and their impact on the metabolism of 7-ethyl-10-hydroxycamptothecin and flavopiridol anticancer drugs. J. Pharmacol. Exp. Ther., 307, 117–128 [DOI] [PubMed] [Google Scholar]
- 58. Seidegård J, et al. (1988). Hereditary differences in the expression of the human glutathione transferase active on trans-stilbene oxide are due to a gene deletion. Proc. Natl. Acad. Sci. U.S.A., 85, 7293–7297 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59. Josephy P.D. (2010). Genetic variations in human glutathione transferase enzymes: significance for pharmacology and toxicology. Hum. Genomics Proteomics, 2010, 876940 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60. Hein D.W. (2006). N-acetyltransferase 2 genetic polymorphism: effects of carcinogen and haplotype on urinary bladder cancer risk. Oncogene, 25, 1649–1658 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.